OpenMP offers five scheduling schemes for the #pragma omp parallel for directive.

However, out of these, only static, dynamic, and guided are the real deal when it comes to scheduling.

The other two, runtime and auto, aren’t true scheduling schemes themselves but just use the others depending on runtime or default settings. They cannot be set with a chunk size, either.

So, it looks like we should focus our testing on static, dynamic, and guided, right?

I think in lecture, it only mentions static, dynamic, guided, runtime. With runtime you can set the chunk size in script. But I am also confused whether I should make runtime as a indepedent strategy.